Google Gemma 3 Beats DeepSeek V3: 100% FREE and Local!
Google Gemma 3 Beats DeepSeek V3: 100% FREE and Local!
# Gemma 3: Most Capable Model You Can Run on a Single GPU or TPU
🔥 **INTRODUCING THE REVOLUTIONARY GEMMA 3** 🔥
Run the most powerful AI model locally with just a single GPU or TPU!
Code: https://mer.vin/2025/03/gemma-3-create-agents/
## 🚀 KEY FEATURES:
– Outperforms Llama 4’s 5 billion parameter model & DeepSeek v3.03 mini in human…
source
Reviews
0 %
I am using BeLikeNative. A simple no-prompt writing assistant that does not generate run-of-the-mill content for my team. Rather, this tool paraphrases and translates non-fiction and fiction content in 11 different languages. The best part is that my team knows barely 3 languages properly. Still, we deliver content, including creative content, with native-level accuracy. So far, no complaints from clients.
Awesome Video. I used AWS EC2 g4dn.xlarge (4vCPU/16G RAM) with Ollama + gemma3:1b, it took 9 seconds to respond to the meal suggestion you did. Does anyone have an idea what model will work faster with this instance?
Just Subscribed to your channel thanks fro the info sir 🙂
Gemma 3 does not out perform deepseek v3 in ANY metric…
@MervinPraison : What is the PIP installed "praisonaiagents[llm]" ?
On which computer were you running Gemma 3? Processor, RAM, GPU?
this doesnt work as well as described. I am running into so many issues
Can i use it locally with a pdf as input?
You have real bad plug for your agent packages at the end. Those are not agents that's pure tomfoolery. 🫡
I am getting error with multiple agent code. I am running on windows 11.
"Error in LLM response: litellm.BafRequestError: Invalid Message passed in {'role': 'system' 'content':'you are nnyour Role : Assistant nn your goal: }
Used it yesterday for a task that required generating SQL queries. Qwen outperformed gemma 4B.
Your concerns about algorithmic bias and potential financial incentives in Google's AI ecosystem are supported by several recent developments. 😂
great demo! you mentioned that you will put all the code in the description below, but i couldn't find it
Sucks at coding, tho.
where is praisionaiagent course?
So now Janus as multimodal is obsolete. It is hard to believe that it can beat much larger models. I hope this is not the hype like the qwq. Too good to be true.
I wonder what kinda AI "they" actually have in the back😮😮, what appears only the surface guys. This is insane
Everyday there is some new model that beats another one. How can anyone keep up?
code? 😊
Some man taught me – the best AI is always free. How he was right!
Gemma 3 has the knowledge of the world to be up around of 2021 year. Not later.